2.3 Study Conduct
The study included four visits: screening, baseline/randomization, end of the first double-blind treatment period/re-randomization into a second double-blind treatment period, and end of the study/early termination. At screening, subjects signed informed consent; had a medical history and physical examination; and were assessed for vital signs, serology, urine pregnancy test, clinical laboratory tests, urinalysis, urine drug test, and 12-lead electrocardiogram (ECG). Hematology, clinical chemistry, urinalysis, urine drug screen, and human immunodeficiency virus, hepatitis B surface antigen, and hepatitis C virus antibody analyses were performed by Eurofins Central Laboratory (Lancaster, PA, USA). For qualified subjects, the investigator or designee contacted the designated interactive web response system to obtain a unique patient identification number for use throughout the study. Subjects were trained on how to enter migraine and study medication data into a real-time electronic diary (eDiary). The eDiary was dispensed at screening for subjects to enter at least one migraine attack to ensure their comfort and compliance with the device. Subjects continued to take their normal migraine medication during the screening period, which lasted up to 21 days to ensure that they met the eligibility criteria. Once eligibility for enrollment was confirmed, the investigator or designee contacted the interactive web response system centralized randomization center to receive the blinded study medication kit assignment.
At baseline, subjects were randomized (1:1) to receive DFN-02 or matching placebo. Study site staff instructed subjects in the proper self-administration of study medication and subjects were confirmed to have, or were provided with, written instructions for use.
In the first double-blind treatment period, subjects were instructed to use one nasal spray device to treat a single migraine attack as soon as (and no more than 1 h after) they experienced migraine pain of moderate to severe intensity. If subjects were unable to use study medication for the first attack after randomization, they were instructed to use the study medication for the next qualifying attack. Subjects were permitted to take a second dose of study medication or rescue medication if they failed to experience sufficient relief from the first dose within 2 h and they had recorded a 2-h postdose assessment in the eDiary. The choice of rescue medication was made by investigators and subjects, and no more than two doses of study medication were permitted within a 24-h period.
Subjects who treated an attack and recorded their response to medication in an eDiary were re-randomized into a second double-blind treatment period in which they treated a second migraine attack at any level of pain intensity. The study was powered and data are presented for the first double-blind treatment period. After they were randomized, subjects could participate in the study for no more than 10 weeks.
Subjects recorded information about their attacks, the effects of study medication on functional disability, and satisfaction with treatment, in the eDiary; other data, such as safety and tolerability, were entered by site staff into electronic case report forms. After study completion or discontinuation, subjects were referred to their usual healthcare professional to continue treatment of migraine and other conditions.
2.4 Assessments
This report presents the results of the DFN-02 comparison with placebo on the efficacy endpoints of functional disability and subject satisfaction with treatment at 2 h and 24 h postdose. Other efficacy data from this study, including data on the endpoints of pain freedom, pain relief, and absence of the MBS, have been presented elsewhere [
21].
Functional disability was assessed at predose, 2 h postdose, and 24 h postdose on a scale where 0 = no disability, able to function normally; 1 = performance of daily activities mildly impaired, could still do everything but with difficulty; 2 = performance of daily activities moderately impaired, unable to do some things; and 3 = performance of daily activities severely impaired, unable to do most things, bed rest may have been necessary [
4]. Satisfaction with treatment at 2 h postdose was evaluated on a 7-point Likert-type scale [
23], where 1 = very satisfied, 2 = satisfied, 3 = somewhat satisfied, 4 = neither satisfied nor dissatisfied, 5 = somewhat dissatisfied, 6 = dissatisfied, and 7 = very dissatisfied. Scores were categorized as satisfied (1–3), satisfied/neither (1–4), neither/dissatisfied (4–7), and dissatisfied (5–7), and the degree of subject satisfaction with treatment was defined by these categories.
The Patient Perception of Migraine Questionnaire-Revised (PPMQ-R) [
24,
25] was used at baseline (pre-randomization) to assess subjects’ satisfaction with their usual pre-study migraine treatment, and was also used at 24 h postdose to assess subjects’ satisfaction with the study medication. The PPMQ-R scale comprised items for efficacy (11 questions), function (4 questions), ease of use (2 questions), and tolerability (10 questions). Cost-related items were excluded because there was no charge for study medication. Three global items measured subject satisfaction in terms of medication effectiveness, adverse effects, and overall satisfaction. The total score was comprised of the efficacy, function, and ease of use subscale scores. The tolerability of adverse effects (i.e. bothersomeness of adverse effects) was scored on a 5-point Likert-type scale that ranged from 1 (not bothersome at all) to 5 (extremely bothersome). Each PPMQ-R subscale score and the total score were transformed to range from 0 to 100 by subtracting the lowest possible scale score, dividing the range of the scale, and multiplying by 100, with higher scores indicating better satisfaction or tolerability. For the transformed scale, higher scores mean greater satisfaction. The total raw score and the three global items were not transformed, and therefore lower scores denote better satisfaction. If a response was missing, the particular subscale or global item was considered non-evaluable. If a subscale or global item was deemed non-evaluable or missing, the corresponding total score was also considered non-evaluable.
Safety variables included the incidence of adverse events (AEs), clinical laboratory data (hematology, chemistry, and urinalysis), vital sign measurements (sitting blood pressure, pulse, respiration rate, and body temperature), physical examination findings, 12-lead ECG readings, urine pregnancy test results, study medication use, and concomitant medication use. The safety endpoint assessments included tolerability as assessed by AEs for each treatment period, and overall and safety as assessed by clinical laboratory tests, vital signs, and ECGs for each treatment period and overall.
Adverse events were coded using the Medical Dictionary for Regulatory Activities (MedDRA, version 19.0) and classified by severity (mild, moderate, severe) and causality (not related, possibly related, probably related, definitely related). Severe AEs were defined as AEs that prevented normal everyday activities and usually needed treatment or other intervention. Investigators also characterized the seriousness of AEs, and serious AEs (SAE) were defined as any untoward medical occurrences or effects that, at any dose, resulted in death or were life-threatening, required or prolonged inpatient hospitalization, resulted in persistent or significant disability/incapacity, or were congenital anomaly/birth defects. Not all severe AEs were SAEs.
2.5 Statistics
The study populations for analysis included all subjects who were screened (for disposition), all randomized subjects who took at least one dose of study medication and had at least one postbaseline efficacy time point (for efficacy), and all subjects who received at least one dose of study medication (for safety). Data presented are based on a statistical analysis plan with prespecified endpoints, except for two data points (percentage of subjects with no disability, and percentage of subjects with mild/no disability), which were analyzed post hoc. Analyses and summaries were produced using SAS version 9.3 or above (SAS Institute, Cary, NC, USA).
Analyses on quantitative and categorical variables included data from subjects with non-missing values, and missing or incomplete data were imputed only for summaries of AEs. Unless specified otherwise, all statistical testing and confidence intervals (CIs) were two-sided and were performed using a significance (alpha) level of 0.05. For analyses of functional disability, the number of subjects with response, the number of subjects with non-missing assessments, proportions for the DFN-02 and placebo groups, 95% exact CIs for those proportions, the odds ratio for response, and the corresponding p values from Fisher’s exact test were computed for comparisons between the treatment groups. For the assessments of subject satisfaction, p values from the Wilcoxon rank-sum and signed-rank tests were computed.